feat: no VRAM plan, the prediction config asks for a measured batch e1b363a verified VBoussot commited on 17 days ago
Build total_mr-3mm's UNet with the five levels M852 was converted to 4bc8591 verified VBoussot commited on 22 days ago
Performance table measured against the original on the same torch and GPU (2026-09-09) becb597 verified VBoussot commited on 29 days ago
Rename ResampleToResolution to Resample, the spelling konfai 1.8 accepts e881cc5 verified VBoussot commited on Aug 7
docs: perf comparison table (time/RAM/VRAM) vs the original tool, S/M/L 4892ed9 verified VBoussot commited on Jul 18
fix(total_mr-3mm): merge multi-model labels with MergeLabels (cumulative offset), not Sum 662c39e verified VBoussot commited on Jul 18
fix(total_mr): merge multi-model labels with MergeLabels (cumulative offset), not Sum 714d9e8 verified VBoussot commited on Jul 18
fix(total-3mm): merge multi-model labels with MergeLabels (cumulative offset), not Sum a15b8bb verified VBoussot commited on Jul 18
fix(total): merge multi-model labels with MergeLabels (cumulative offset), not Sum 6e040e7 verified VBoussot commited on Jul 18
perf: recalibrate vram_plan (leave GPU headroom) + refresh perf README 27bb488 verified VBoussot commited on Jul 6
perf: recalibrate vram_plan (leave GPU headroom) + refresh perf README 333ce14 verified VBoussot commited on Jul 6
Fix model-intrinsic params in the model (input channels / class count / network architecture) instead of exposing them in the config 769f01c verified VBoussot commited on Jul 5
app.json: drop redundant top-level patch_size (now read from Prediction.yml) 9274d29 verified VBoussot commited on Jul 5
Update model card, app.json (patch size + VRAM plan) and inference configs f1cd0d2 verified VBoussot commited on Jul 3